Be first to read the latest tech news, Industry Leader's Insights, and CIO interviews of medium and large enterprises exclusively from Medical Tech Outlook
THANK YOU FOR SUBSCRIBING
By
MedTech Outlook | Tuesday, May 30, 2023
The ethical implications of artificial intelligence (AI) systems include the taxonomy of algorithmic bias and the implications of fairness in the design of AI systems.
FREMONT, CA: The convergence of AI and interventional and surgical specialities will generate a slew of ethical concerns. The majority of these are about bias and accountability. Moving towards a fully autonomous manner of functioning will intensify ethical concerns. This is especially crucial for self-directed AI-initiated responses and interventions. The datasets used to train AI may already be biased, as was covered in-depth in another section. This will raise ethical issues, particularly if different patient categories are affected differently downstream. This could imply that additional interventions have varying effects on certain patients.
Stay ahead of the industry with exclusive feature stories on the top companies, expert insights and the latest news delivered straight to your inbox. Subscribe today.
AI can now estimate a person's age by identifying the facial characteristics that may have helped in that assessment. This in turn aids in formulating surgical techniques that, by changing certain traits, could delay the appearance of age. Motion-sensor surgical tools are used by surgeons to collect data in real time and direct the surgeon to make minute modifications to improve the result. However, there are some biases built into these AI algorithms.
The stakes are much higher in a surgical setting or a procedural laboratory. A fatal result could occur if an AI-trained robot dissecting, suturing, or manipulating catheters inside the heart freezes due to a technical glitch or loses control during the treatment. The ethical concerns will directly relate to how much AI is used. It would be crucial to train robots using datasets of thousands of processes carried out under various conditions at several sites with varied operators.
The decisions surgeons make based on instinct cannot be duplicated by AI. Since most of the gut guidance reflex comes from unquantifiable clinical experience, it is difficult to capture and replace. Additionally, a single surgical procedure entails hundreds of complex processes, including cutting, dissecting, excising, joining, burning, chilling, clamping, ligating, and suturing. For the foreseeable future, robots will only be used to assist as they get more adept at performing simple tasks, new layers of complexity will be carefully introduced.
Although recent developments like the Smart Tissue Autonomous Robot (STAR) have started to perform surgical tasks more frequently, applications of AI in surgery are still largely limited to the agents performing specific and defined tasks initiated and controlled entirely by human surgeons or clinicians albeit under very particular and regulated circumstances. Before surgery or other acute care, AI may also be employed as a clinical decision support system. In such situations, de-ontological ethics may be used to completely encapsulate the ethical paradigm. According to this theory, an AI system is an implicitly moral agent, and it is the human designers and developers of the machine agent who are responsible for the behaviour of the agent.
An AI system's capacity for decision-making results from an underlying mapping between pertinent inputs and an output choice. No matter the learning mechanism, confidential patient data is typically needed to train a production-grade AI system for surgical applications to maintain the patient's confidentiality. Since it may be claimed that personally identifiable medical information is the most sensitive type of user data now available, it is consequently an automatic topic of discussion in many conversations regarding the ethics of AI in surgery. Therefore, while using AI in surgery, the protection of users' medical data is of the utmost importance and a top priority.
Each medical professional who will be impacted by the incorporation of AI in surgery will be involved in monitoring it. In particular, the Surgeon-in-Chief can represent the institutional memory and authority to make sure that advancement is handled effectively and securely, in cooperation with a committee of stakeholders. It appears that human designers and developers are ultimately responsible for a machine's behaviour. Bridging the gap between AI researchers and the various stakeholders (or the end users) of the systems they develop is an efficient way to deal with eventualities.
Innovation in surgery is a process rather than an event, and it frequently comes about as a result of original solutions to specific issues. Through a process of careful consideration, it will be possible to separate good ideas from terrible ones. The most creative ideas may not, however, have a short-term influence on care if there are too many filters because they may be too complex.
Ethical considerations of using AI in surgery frequently involve complying with data protection and privacy laws, fairness (broadly defined) throughout data source, development, and deployment, and the upholding of specific transparency norms. Machine learning, like ethical decision-making in present practice, will not be effective if it is only well constructed by committee - it requires exposure to the real world.
More in News
I agree We use cookies on this website to enhance your user experience. By clicking any link on this page you are giving your consent for us to set cookies. More info
